Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix benchmark dir and add CUDA_DEVICE_MAX_CONNECTIONS to qwen #8678

Merged
merged 17 commits into from
Jun 28, 2024
Merged

fix benchmark dir and add CUDA_DEVICE_MAX_CONNECTIONS to qwen #8678

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
ed0ed78
remove tsinghua pypi
fightfat Jun 1, 2023
3ec64f9
Merge branch 'PaddlePaddle:develop' into develop
fightfat Jun 6, 2023
f7558ac
modify gpt dateset addr for benchmark
fightfat Jun 6, 2023
68dce6c
Merge branch 'PaddlePaddle:develop' into develop
fightfat May 23, 2024
cc05e32
fix run_benchmark for llama2_70b in auto_parallel
fightfat May 23, 2024
b761d9c
Merge branch 'PaddlePaddle:develop' into develop
fightfat Jun 12, 2024
99ea6c2
fix autotunner benchmark error and fix llama2 dy2st benchmark
fightfat Jun 12, 2024
839ff62
Merge branch 'PaddlePaddle:develop' into develop
fightfat Jun 24, 2024
e693341
fix benchmark dir because of PR#8627
fightfat Jun 24, 2024
fe43837
Update run_benchmark.sh
fightfat Jun 24, 2024
ed30855
fix benchmark dir because of PR#8627
fightfat Jun 25, 2024
4fe341c
Merge branch 'develop' of github.com:fightfat/PaddleNLP into develop
fightfat Jun 25, 2024
5710fc8
add qwen N4C32
fightfat Jun 25, 2024
f7c8567
fix etcd
fightfat Jun 25, 2024
8f15694
fix qwen N4C32 for per_device_type
fightfat Jun 26, 2024
7425dd1
merge conflict
fightfat Jun 28, 2024
726bfd2
fix benchmark dir and add CUDA_DEVICE_MAX_CONNECTIONS to qwen
fightfat Jun 28, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion tests/test_tipc/dygraph/hybrid_parallelism/ce_gpt/benchmark_common/run_benchmark.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -203,7 +203,7 @@ export FLAGS_cudnn_deterministic=True
export FLAGS_cudnn_deterministic=1
export FLAGS_embedding_deterministic=1

export PYTHONPATH="../../../PaddleNLP/"
export PYTHONPATH="../../PaddleNLP/"
source ${BENCHMARK_ROOT}/scripts/run_model.sh # 在该脚本中会对符合benchmark规范的log使用analysis.py 脚本进行性能数据解析;如果不联调只想要产出训练log可以注掉本行,提交时需打开
_set_params $@
#_train # 如果只产出训练log,不解析,可取消注释
Expand Down
1 change: 1 addition & 0 deletions tests/test_tipc/dygraph/hybrid_parallelism/qwen/benchmark_common/run_benchmark.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -230,6 +230,7 @@ function _train(){
}

export PYTHONPATH=$(dirname "$PWD"):$PYTHONPATH
export CUDA_DEVICE_MAX_CONNECTIONS=1

source ${BENCHMARK_ROOT}/scripts/run_model.sh # 在该脚本中会对符合benchmark规范的log使用analysis.py 脚本进行性能数据解析;如果不联调只想要产出训练log可以注掉本行,提交时需打开
_set_params $@
Expand Down
2 changes: 1 addition & 1 deletion tests/test_tipc/dygraph/moe/gpt/benchmark_common/prepare.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,4 +25,4 @@ python3 -m pip install --upgrade paddlenlp
cd ../legacy/model_zoo/moe/dygraph/
rm -rf data
mkdir data && cd data
wget https://bj.bcebos.com/paddlenlp/models/transformers/gpt2/train.data.json_ids.npz
wget https://bj.bcebos.com/paddlenlp/models/transformers/gpt2/train.data.json_ids.npz
Loading